Replication of retroviruses such as HIV involves two distinctive steps early after infection: reverse transcription of the viral RNA genome to make a DNA copy, and integration of that DNA copy into a chromosome of the host. The choice of host sites for integration is not specific with respect to the target DNA sequence. The lack of integration specificity obstructs many of the otherwise promising applications of retrovirus-based technology in the prevention and treatment of AIDS. For example, an attenuated HIV vaccine might be effective in preventing HIV infection, but safety concerns related to insertional mutagenesis during integration obstruct the development of such a vaccine. Similar safety concerns arise in treating HIV-infected patients by retrovirus-based gene therapy. To overcome these concerns, we propose to develop retroviral derivatives capable of integrating into preselected target sites. We have previously demonstrated in vitro that a fusion protein composed of the HIV integrase protein linked to a site-specific DNA-binding domain directed integration selectively to target DNAs containing sites recognized by the new DNA binding domain. We propose to create replication-competent HIV derivatives that integrate at predetermined sites in vivo by replacing the integrase coding region with sequences encoding fusions of integrase to new DNA binding domains. First, we will construct further fusions of integrase to DNA-binding domains and characterize them in vitro to identify optimal combinations. We will then substitute sequences encoding integrase fusions for the normal integrase coding region in replication competent proviral clones. We will transfect these modified proviruses into cultured cells, harvest viral supernatants, and check particle assembly, protein composition and replication competence. We have already found that an HIV derivative containing new sequences fused to the carboxyl-terminus of integrase is able to replicate, supporting the plausibility of our plan. As an initial test of targeting in vivo, we will purify preintegration complexes from cells infected with our modified retroviruses and assess their ability to target integration in vitro. Only after mutant viruses have scored positive in this assay will we undertake the more laborious task of assessing integration site selection in vivo. A PCR assay will be used to examine integration sites used in various chromosomal regions, and integration sites will be cloned and sequenced using a novel PCR-based strategy. Initially we will attempt to direct integration into repetitive DNA sequences in vivo. These are attractive integration sites since they can potentially accommodate insertions without disrupting required functions, thereby reducing the chances of insertional mutagenesis during vaccination or gene therapy. Longer term, we will aim to direct integration to single copy DNA sequences. Such a system might permit the inactivation of integrated HIV genomes by targeted insertional mutagenesis in patients.